Photochemical transformations of unsaturated bicyclic ketones

Pedro Joseph-Nathan , Myriam Meléndez-Rodríguez , Carlos M. ... Fina Liotta , Robert A. Batey , Gavin D. Heffernan , David E. Cladingboel , Robert C...
1 downloads 0 Views 2MB Size
3828 manually controlled gas buret. Potassium superoxide was ana1yzed38 by the addition of diethyl phthalate (50 vol. %) followed by an equal volume of acetic acid at 0". A vigorous evolution of oxygen occurred. The temperature was brought to room temperature and after shaking for 100 min the oxygen evolution was measured. The yield of potassium superoxide was calculated from the equation 2K02

+ 2CHSC02H +2CH3C02K + H202 +

Azobenzene was isolated from the oxidates after acidification with hydrochloric acid by the addition of 300 ml of water for 20 ml of the initial oxidate. After standing for 24 hr, the precipitated azobenzene was removed by filtration. The crystals were dissolved in methylene chloride and transferred to a tared flask. The solvent was removed under vacuum and the azobenzene, mp 67-68', evacuated for 4 hr before weighing. Recovery tests indicated a loss of 20 mg of azobenzene in this procedure, and this correction has been applied to yields given in the text.

0 2

Photochemical Transformations of Unsaturated Bicyclic Ketones. Verbenone and Its Photodynamic Products of Ultraviolet Irradiation' William F. Erman

Contributionf i o m The Procter & Gamble Company, Miami Valley Laboratories, Ciiicinizati, Ohio 45239. Received August 15, 1966 Abstract: Irradiation of verbenone (1) in cyclohexane or acetic acid with alight source with principal emission >300 mp or a Pyrex filter afforded chrysanthenone (2) of high optical purity as the principal product (37-67 %). In contrast, irradiation of 1 in cyclohexane with a broad-spectrum mercury arc lamp using a quartz or Vycor filter afforded as many as six products: the two decarbonylation products 9 (1-2z) and 10 (6-13 %) and the rearranged ketones 8 (9-20%), 2 (6-16%), l l a (4-6%), and 12 (2-3Z). Three other ketones, 13, 14, and 15, were produced in small amounts. Each of the compounds 2, 8, and 12, from the latter irradiation, though optically active, showed appreciable loss of activity depending upon the length and conditions of irradiation. When alcohols were employed as solvent for the irradiation of 1 using the broad-spectrum lamp, the major products were the ketone 2 (12-20z) and the ester 3 (21 Irradiation of 2 or 8 with the broad-spectrum lamp led to a photostationary state in which the the ratio of 2 to 8 was approximately 1.5: 1. The ketone 14 on exposure to ultraviolet light in methanol solution afforded the three acetals 19 (3 20 (2%), and 21 (11 %) and the ester 3 (5 %). Irradiation of 14 in cyclohexane afforded the hydrocarbon 10 (12%) as the principal isolable product and only trace quantities of 2 (1.8%) and 8 (2.1 %). The above results are discussed in terms of the dual mechanisms proposed by Hurst and Whitham for the formation of chrysanthenone (2) from verbenone (1).

z).

z),

T

he photochemical conversion of verbenone (1) to chrysanthenone (2)2 is one of two recorded examples in which a$-unsaturated ketones undere o photoinduced rearrangement resulting in a n ultimate 1,3-alkyl shift.3 Whitham and Hurst, in their investigation of this transformation, showed that ultraviolet irradiation of verbenone in cyclohexane afforded chrysanthenone with partial optical retention while irradiation in the presence of alcohol or amines yielded, besides ketone 2, derivatives related to geranic acid (e.g., the ester 3b or 4b4 and the esters 5b and 6b resulted from irradiation of 1 in ethanol). They further showed that irradiation of chrysanthenone 2 in ethanol produced the same esters 3b, 4b, 5b, and 6b, while irradiation of optically active 2 for long periods in inert media afforded 2 with considerable loss of optical activity depending upon the period of irradiation.

(1) Paper presented in part before the Division of Organic Chemistry, 152nd National Meeting of the American Chemical Society, New York, N. Y . , Sept 1966, Abstracts, p 29s. ( 2 ) J. J. Hurst and G. H. Whitham, J . Chem. SOC., 2864 (1960). (3) A second example has been provided recently by Zimmerman and Sam, who reported the production of 2-(cis-styryl)-3-phenylcyclobutanone in 8-9 % yield from the ultraviolet irradiation of 4,Sdiphenylcyclohexenone in [-butyl alcohol: H. E. Zimmerman and D. J. Sam, J . A m . Chem. Soc., 88, 4905 (1966). (4) Although two geometric isomers of the &?-unsaturated ester, Le., 3 and 4, could be produced in this reaction, Hurst and Whitham did not separate the isomers or assign stereochemistry to the predominant isomer.

All of these observations were consistent with a twopath mechanism for the formation of chrysanthenone (Chart I): path a, a fast process involving an ultimate 1,3-alkyl migration, and path b, cleavage of 1 to the ketene 7, a n optically inactive species, which undergoes a n intramolecular thermal or photolytic cyclization to 2. Photochemical racemization of chrysanthenone by cleavage t o ketene 7 and recyclization to 2 (path c) would account for part of the loss in optical activity of 2 during the irradiation process. I n an over-all program directed toward the utilization of this reaction in sesquiterpene synthesis, we reinvestigated the photochemistry of verbenone with the hope of (1) effecting the rearrangement with minimal racemization, and (2) discovering other products of the irradiation (by varying reaction conditions) which might shed further light on the mechanism of product formation. Results

(-)-Verbenone (1) was irradiated under a variety of conditions using light sources with broad spectrum emission in quartz or Vycor vessels, using light sources with limited emission in the 280-300-mp region, and using filters which diminish transmission of light below 300 mp. The results of these experiments are listed in Table I.

Journal o f the American Chemical Society J 89:15 1 July 19, 1967

3829 Table I. Irradiation of (-)-Verbenonqa Run

Lamp

1

450 Hanovia 2 450 Hanovia 3 450 Hanovia 4 450 Hanovia 5b 450 Hanovia 6 450 Hanovia 7 450 Hanovia 8 450 Hanovia 9 450 Hanovia 10 450 Hanovia l l b 450 Hanovia 12b 450 Hanovia 13 2537A Rayonet 14 3500A Rayonet 15 3500A Rayonet

Time, Temp, hr "C Jacket

[.Iz6D

-256"

Solvent

Z unreacted 1

9

Cyclohexane

0.4

1.5

Yield, % 8 2

[aIz6D, deg

7

10

14

13

15

3

9

0.1

0.2

20 16

4

2

0.4

...

6

0.3

0.5

1911

6

3

0.5

. . . . . . . . . . . .

. . . +14 -16 . . . . . . . . . -9'... ............

11

12

2

7

1.0

20 Vycor

1.0

20 Quartz Cyclohexane Trace 1

3.0

20 Quartz

Cyclohexane

0

2

12

0.1

0.1

13

9

4

2

1.0

3.0

70 Quartz Cyclohexane

0

2

13

1.1

0.8

9

6

3

2

1.0

1.0

20 Vycor

Benzene

0

0.8

5

Trace Trace

14 10

2

1

0

1.0

20 Pyrex

Cyclohexane

0

0.1

0.4 Trace Trace

7 54

3

1.5 0

1.0

20 Pyrex

Aceticacid

0

1.4

1.0 0

...... ......

0.16 20 Vycor

Aceticacid

17

0.5

20 Vycor

Aceticacid

0

1.0

20 Vycor

Aceticacid

0

1

5

0.5

20 Vycor

Methanol

1

0

0

1.0

20 Vycor

Methanol

0

0

0

0.3

0.9

6.0

27 Quartz Cyclohexane

8

0

0

0

2

6.0

27 Quartz Cyclohexane 10

0

0

0

Trace

2 38 Trace 2

6.0

27 Quartz Aceticacid

0

0

0

Trace

2 43 Trace 0

5

Trace Trace Trace Trace 0

0

Trace Trace

5

67

+24

-28

13

+19

-36

...

-76

... ...

0.3

0.5

2 3 8

0

0.50

. . . . . . -69

0.5

0.3

8 42

1

0

0

0.5

0.3

2315

1

0

0

Trace Trace

4 21

2

0.2 0

3

0

0

. . . . . . . . . . . . . . . . . . . . . . . . 21 . . . . . . . . . 23 . . . . . . . . . . . . . . . -52 +41 . . . . . . -50 . . .

0

. . . . . . -85

10

8

0

7 42 Trace 7

Trace

...

a Solutions of 1.50 g of verbenone (1) in 150 ml of solvent were irradiated in runs 1-12;solutions of 1.00g of 1 in 100 ml of solvent were irradiated in runs 13-15. b 1,5-Dimethyl-5-carbomethoxymethylcyclohex-l-ene (16) was also produced from run 11 in 2-3 yield and from run 12 in 4 yield, An unidentified ester was also produced in 2-3 yield from each run, 1 1 and 12. c The difference in rotation of 2 at 70 and 20" may be explained by the fact that chrysanthenone undergoes slow thermal isomerization at these temperatures.

Chart I

en-6-one (chrysanthenone, 2), (+)-2,4,4-trimethylbicyclo[3.l,l]hept-2-en-6-one (€9, 2,4,4-trimethylbicyclo[3.1.O]hex-Zene (9), 2,6,6-trimethylbicycIo[3.1.O]hex-2ene (lo), (+)- 1,2-dimethyltricyclo[3.3.0.0zJ]octan-6-one (lla), and (+)-isopiperitenone (12) (Chart 11). Three Chart I1 path b

\hr

COOR

6a, R=Me

b,R=Et

L

ROH

$% 7 h.11

5a, R=Me

b, R=Et

10

(+)-I2

3a, R =Me 4a, R=Me

b,R=Et

b, R = E t

When a broad-spectrum, 450-w Hanovia mercury arc lamp and an inert solvent, cyclohexane, were employed for these irradiations, as many as six products (2, 8-12) were isolated even when we attempted to duplicate the conditions of Hurst and Whitham (i-e., run 4, Table I):5 (-)-2,7,7-trimethylbicyclo[3.l.l]hept-2( 5 ) The apparent anomalies between our results and those of Hurst and Whitham are, to a large extent, explained by differences in spectral properties of the specific lamps employed. Indeed, with our system,

13

14

15

other compounds were formed in small quantities: 2,4,4-trimethylbicyclo[3.2.0]hex-2-en-7-one (13), 2,6,6complete consumption of 1.25 g of verbenone as a 1 % solution i n cyclohexane was observed after an irradiation period of less than 30 min compared to the 3-hr period reported by Hurst and Whitham. Exactly the same quantities of verbenone (1.25 g) and cyclohexane (125 ml) were employed in our initial runs in order to compare results with those of Hurst and Whitham. Subsequent experiments listed in Table I employed 1.5 g of verbenone in 150 ml of cyclohexane.

Erman

1 Ultraciolet Irradiation

of Verbenone

3830

triniethylbicyclo[3.2.0]hex-2-en-7-one(14), and piperitenone (15). An infrared spectrum of the crude reaction product showed the presence of ketene (4.72 p) even when the irradiation was terminated after 5 min. The ketene could not be isolated as such from the reaction mixture but was trapped as the ester 3a by addition of methanol to the reaction mixture after irradiation was complete. The compounds 2 and 8-12 accounted for 30-7475 of the product in each run while polymer and possibly ketene dimers accounted for the remaining 26-70%. Each of the ketones 2 and 8 from irradiations with the broad spectrum lamp, though optically active, showed appreciable loss of activity depending upon the period or irradiation. Thus irradiation of verbenone, [ Q ] * ~ D -256", for a period of 1 hr (run 1) produced ketone 2, [ Q ] * ~ D-28", and ketone 8, [ a ]+24" ~ (27% retention), while irradiation for a period of 3 hr yielded 2 and 7 with [ a I z 6-16 ~ and $14' (15% retention), respectively. The isopiperitenone (12) isolated from the same 1-hr run was racemized to about the same extent as ketones 3 and 7, [ a ] +19" ~ (23% retention). Surprisingly, the tricyclic ketone l l a from run 1, [ a I z 6 D 158" (61 % retention), showed considerably greater retention of optical activity than its immediate precursor, isopiperitenone (12).6 In contrast to the above experiments with the broadspectrum lamp, chrysanthenone (2) was almost the exclusive product when a light source with principal emission in the 2537- or 3500-A region' or a Pyrex vessel which filters >90 % of light of wavelength below 300 mp was employed for the irradiations. The rearrangement proceeded with somewhat greater optical retention in the latter instances (compare runs 6 , 13, and 14 to run 1, Table I). When the irradiation was performed in methanol (runs 11 and 12) the principal products were chrysanthenone (2, 8-21 %) and the seqcis-P,y-unsaturated ester 3 (21-23%).8 A small quantity (2-3%) of a second ester, 1,5-dimethyl-5-carbomethoxymethylcyclohex-1ene (16),9was isolated from irradiation of 1 in methanol, but not even trace quantities of the trans-P,y-unsaturated ester 4a and the conjugated geranate 5a and nerate 6a esters were observed in the product mixture.'O When the irradiations were performed in acetic acid using a Pyrex filter or the 3500-A lamps, chrysanthenone of high optical purity relatively free of side products was isolated in 76 and 43% yields, respectively. That acetic acid is the preferred solvent for limiting racemization is indicated by comparing the rotation of ketone 2 isolated from run 6 ( [ a ] D -36",

+

(6) If the rearrangements proceeded with complete optical integrity the products would have been expected to possess the following estimated optical rotations (see Configurational Assignments): 2, -108"; 8, -105"; l l a , -254"; 12, -83". (7) For these experiments a Rayonet photochemical reactor was employed using lamps with principal emission in the region of 2537 and 3500 A, respectively. With the former, bands of about 0.01 the intensity of the principal emission occur in the 280-300-mp or n K* region of chrysanthenone (296 mp). The 3500-A lamp shows principal intensity above 310 mp with maximum emission at 350 mp. (8) The seqcis and seqtrans terminology are used in accordance with the rules suggested by R. S . Cahn, C. Ingold, and V. Prelog, Angew. Chem. Intern. Ed. Engl., 5, 385 (1966). (9) Structure proof of this ester will be presented in a separate communication by W. F. Erman and T. W. Gibson, to be submitted. (10) Hurst and Whitham isolated, as the major product, ethyl geranate (3), as well as the nonconjugated isomer 4b, from the irradiation of verbenone in ethanol. However, irradiations in the presence of amines produced only the nonconjugated isomers. --t

Journal o j t h e American Chemical Society

cyclohexane) and run 7 ( [ a ]-76", ~ acetic acid). It should be emphasized that temperature is a critical factor in ruling the course of reaction in acetic acid. For example, when a 33% solution of (+)-verbenone (+253") in acetic acid was irradiated at 40-50" for a period of 66 hr with the 3500-A lamp, the ketone 14, was a major product (22%),11 while [ a I z b-210", ~ the chrysanthenone (30 %) was highly racemized, [aIz5D +14"." In order to determine the role played by the ketones 2, 8, 12, 14, and 15 in the production or photochemical equilibrium of products, each was exposed to ultraviolet irradiation with the 450-w mercury arc lamp. Chrysanthenone (2, 273 mg) when irradiated for a period of 20 min produced four major products: 9 (273, 10 (13%), 8 (52%), and 2 (33%). When the irradiation was terminated after a shorter period, the ratio of ketone 2 to 8 was greater; however, longer periods of exposure did not alter this ratio appreciably. Irradiation of the ketone 8 for periods identical with that used for chrysanthenone afforded the same products in almost identical yield (40%) and relative composition: 9 (473, 10 (13%), 8 (56%), and 2 (27%). Further irradiation gave little change in ratio of products but considerable decrease in yields. Thus an apparent photostationary state between 8 and 2 is reached after about 20-min irradiation. Our many photochemical studies on verbenone suggest that the ratio of 8 :2 is approximately 1.5-1.6 : 1 when this photostationary state is attained (see Table I),1z If ketene 7 were involved as an intermediate in the photochemical transformation of verbenone to products, one might anticipate that optically inactive ketone 14 would be produced in significant quantities from this reaction. Since racemic 14 was not produced photochemically in significant quantity from any of these runs, we considered the possibility that 14, though formed initially, was converted rapidly to the other observed products of the reaction, For this reason the photochemical behavior of 14 was investigated in some detail. Based on documented examples of the photolytic behavior of 7,7-dimethylbicyclo[3.2.0]hex-2-en-6-one13bor -bicyclo[3.2.0]hex-2-en-7-one,13a~c one might predict that the ketone 14 could undergo photochemical cleavage cia either or both paths a or b (Chart 111). If path b proceeded at a rate faster than a, one might further predict that the ketone 14, once formed from verbenone, would be in a photodynamic equilibrium with chrysanthenone 2 cia the ketene intermediate 7. In order to determine if the ketene 7 was produced to any extent, the irradiation of 14 was first performed in methanol in order to trap at least a minimal quantity of the ketene as the ester derivative 3a. (11) It should be emphasized that chrysanthenone (2) is converted to ketone 14 on treatment with acetic acid under essentially the same conditions in the absence of light. We have also found that chrysanthenone undergoes thermal racemization on extended exposure to temperatures employed here. A description of the photochemical, thermal, and acid-catalyzed rearrangements of chrysanthenone and the mechanistic implications with regard to the photochemical rearrangement of verbenone will be reserved for a separate communication by W. F. Erman, to be published. (12) For a further discussion of this and related photochemical reactions of B,r-unsaturated ketones, see W. F. Erman and H. C. Kretschmar, J . A m . Chem. SOC.,89, 3842 (1967). (13) (a) D. I. Schuster, M. Axelrod, and J. Auerbach, Tetrahedron Letters, 1911 (1963); (b) H. U. Hostettler, ibid., 687 (1965); (c) G. 0. Schenck and R . Steinmetz, Chem. Ber., 96, 520 (1963).

1 89:15 / July 19, 1967

3831 Chart III Irradiation of 14 in methanol, in fact, did produce the cis-P,y-unsaturated ester 3a in small but significant yield (5%). (As in the photolysis of verbenone, only the cis isomer was formed; not even trace quantities of the trans ester 4a, nerate ester 6a, or geranate ester . I H ’ H Sa were observed in the reaction mixture.) The principal products of the reaction, however, were an alde19 CH,OH hyde (14%), which we were unable to identify because 2o of its apparent lability, and the three acetals, 2,6,6trimethyl-7-endo-methoxy-8-oxabicyclo[3.3.0]oct - 2- e n e 8_ ,@$ -co(19, 3 %), 2,6,6-trimethyl-7-exo-methoxy-8-oxabicyclo[3.3.0]oct-2-ene (20, 2 %), and 2,6,6-trimethyl-8-exo10 14 17 methoxy-7-oxabicyclo[3.3.0]oct-2-ene(21, 11 %). l 4 Irradiation of 14 in cyclohexane afforded only trace “0 quantities of the two ketones 8 (2.197) and 2 (1.8%) (which could arise by intramolecular cyclization of the ketene 7), along with a number of minor products including the unidentified aldehyde 22 (1.1 %). The principal isolable product, however, was the hydro18 8 carbon 10 (1297). The total absence from the reaction 21 mixture of the hydrocarbon 9, trace amounts of which could be detected easily by gas chromatography, certainly confirms that the latter hydrocarbon 9 is not formed uia 10 (see Discussion) in the irradiation of verbenone. Even when the irradiation was performed in benzene as a photosensitizer, the cyclopropane 9 u 7 3a was not detected in the product mixture. 2 It seemed obvious at first glance that isopiperitenone of acyclic a,P-unsaturated ketones t o the P,y position’g (12) was the photochemical precursor of the ketone prompted us to consider the possibility of photocheml l a . Indeed, irradiation of racemic isopiperitenone in ical rearrangement of piperitenone (15) to isopipericyclohexane with a broad-spectrum lamp and Vycor tenone (12). If such a photochemical conversion did filter did produce as one product l l a but in very poor occur, then this would account for the absence of On the other hand, irradiation of 12 yield (“3%). piperitenone in these runs and also make available a using a Pyrex filter afforded l l a in 36 % yield. I6 convenient synthesis of isopiperitenone. UnfortuIt is quite surprising that the ketone 15 is absent nately, irradiation of piperitenone (15) for as long as from many runs when 12 is present in significant quan6.5 hr in cyclohexane or ethanol afforded only polytities.Is The reports of Yang on the isomerization meric materials and unchanged 12. *O Structure Proof of Products. The structures of the (14) We were unable to detect even trace quantities of methyl amethylpropionate (ii) which would have arisen from the anticipated two hydrocarbons 9 and 10 were easily deciphered by ketene fragmentation product i of the diradical 17.13 infrared and nmr spectroscopy. The infrared spectrum of 9 showed absorption at 6.1 1 and 12.0 p typical of a trisubstituted olefin and at 9.75 p, a characteristic band for cyclopropane.*’ The nmr spectrum was well resolved in benzene (100 Mc). Since chemical shifts i ii and spin-coupling constants are recorded in the Experi(15) Prinzbach and Druckrey, on the other hand, have found that the mental Section, only two aspects of the spectrum closely related cyclopropane iii is transformed cleanly to iv on ultrawhich clearly define the structure as 9 will be discussed violet irradiation. The ester iii is thermally more stable, however, since iv is converted back to iii at 220”: H. Prinzbach and E. Druckrey, here. Since the endo-cyclopropyl methylene proton Tetrahedron Letters, 2959 (1965). lies directly over the T orbitals of the olefin, a strong diamagnetic chemical shiftz2 of this proton to r 9.74 is observed. The exo proton is well separated from hv the endo at 7 9.39. That these are the correct assign7 ments for the endo and exo protons is confirmed by the spin-coupling patterns of each with the H-1 and H-5 iii iv

2

cH30Y go &

-

I

tl

- 8

(16) It also should be emphasized that the analogous photoinduced cyclization of carvone to carvonecamphor is best performed in ethanol as solvent using sunlight1’ or a light source with principal emission above 310 mp. We are investigating the photochemical behavior of isopiperitenone in further detail. These studies, including irradiations in alcohol solvents, will be reported in a separate communication.g (17) (a) G. Buchi and I. M. Goldman, J. A m . Chem. Soc., 79, 4741 (1957); (b) J. Meinwald and R. A. Schneider, ibid., 87, 5218 (1965); (c) T. W. Gibson and W. F. Erman, J . Org. Chem., 31, 3028 (1966). (18) In fact, piperitenone may well be formed from isopiperitenone (12) during work-up exclusively. Thus 12 is converted to 15 by mild acid or base conditions including column chromatography separations on alumina or silica gel.

(19) N. C. Yang and M. J. Jorgenson, Tetrahedron Letters, 1203 (1964). (20) Photocyclodimerization of ethanolic or acetic acid solutions of the closely related piperitone occurs in sunlight [W. Treibs, Chem. Ber., 63, 2738 (1930); H. Ziffer, N. E. Sharpless, and R. 0. Kan, Tetrahedron, 22, 301 1 (1966)]. Indeed, photodimerization may be the principal course of reaction with piperitenone. (21) Unless otherwise stated, infrared interpretations are derived from L. J. Bellamy, “The Infra-red Spectra of Complex Molecules,” 2nd ed, John Wiley and Sons, Inc., New York, N. Y.,1958. (22) L. M. Jackman, “Applications of Nuclear Magnetic Resonance Spectroscopy in Organic Chemistry,” Pergamon Press, New York, N. Y., 1959, p 129.

Errnan J UItrauioIet Irradiation of Verbenone

3832

protons.23 The exo proton is cis t o the H-1 and H-5 protons and should exhibit greater coupling (6-8 cps) than the endo proton trans t o H-1 and H-5 (3-5 CPS).*~ Indeed, the methylene coupling pattern was that of an ABX2 system with J (H-6-end0, H-6-exo) E 4.0 cps, J (H-1 and -5, H-6-exo) g 6.0 cps, and J (H-1 and -5, H-6-endo) g 3.0 cps, as anticipated. One other distinguishing feature of the nmr spectrum of 9 is the vinyl C-3 proton signal which displays coupling of less than 1 cps. In contrast the vinyl C-3 proton exhibits considerably greater splitting in the nmr spectrum of 10. This observation is consistent with the presence of the gem-dimethyl function at C-4 in 9 and a methylene at C-4 in structure 10. The same difference in chemical shift for the endoand exo,gem-dimethyl protons in the nmr spectrum of 10 is observed as for the methylene protons of 9. This observation plus the well-defined but broadened AB quartet exhibited by the allyl protons (centered at T 7.85) serve t o define the second hydrocarbon 10. The infrared, nmr, and mass spectra of compounds 13 and 14 were strikingly similar. The two compounds were resolved only with difficulty by gas chromatography. Since the structure of 14 was readily identified by comparison of the nmr, mass, infrared, and ultraviolet spectra with an authentic sample of 14 prepared by the method of Beereboom,2j we were encouraged t o propose the structure 13 for the second compound. The compound 13 showed strong absorption in the infrared spectrum for a cyclobutanone at 5.62 p , 2 4 a gem-dimethyl doublet at 7.25 p , 2 1and a trisubstituted olefin at 6.1 1 and 11.9 The distinguishing feature of the nmr spectra of both 14 and 13 was the heavily coupled C-1 proton which occurred at T 6.0 in both isomers. Final confirmation of structure 13 was presented by analysis of the nmr spectrum and spin decoupling experiments. The C-4 gem-dimethyl and C-2 vinyl methyl signals were readily apparent and are detailed in the Experimental Section. The C-6 protons appear as part of an ABXY system with the AB portion centered at 7 6.98 with JAB= 17.6 cps, JBX= 8.1 cps, JAx = 6.0 cps, J B y = 4.0 cps, and JAY = 2.4 cps as anticipated for a slightly puckered four-membered ring. The C-5 (H,) proton appears as a multiplet centered at T 7.44. However, irradiation of the C-1 (HY) proton with a frequency of 2717 cps resolved the Hx spectrum into a simple quartet of JBX= 8.1 cps and J A X = 6.0 cps. 2 G

edly similar. Since comparison of the infrared and ultraviolet spectra of the first isomer with that published for natural c h r y ~ a n t h e n o n eisolated ~~ from the essential oil Chrysanthenium indium established the structure 2 for this ketone, we were prompted t o assign structure 8 t o the second isomer. Confirmation of structure 8 is apparent from the following arguments. The infrared spectrum of 8, like that of chrysanthenone, showed strong carbonyl absorption at 5.62 p suggestive of a cyclobutanone.21 The ultraviolet spectrum was that of a typical P,runsaturated ketone with A,, 297 mp (E 180). Both ketones 2 and 8 were cleaved by alcoholic potassium hydroxide t o a mixture of the two carboxylic acids, 2,2,4trimethylcyclohex-3-enylcarboxylicacid (23) and 2,2,4 trimethylcyclohex-4-enylcarboxylic acid (24); the former (2) afforded 23 and 24 in the ratio 18: 13 (55% yield), while the latter (8) gave 23 and 24 in the ratio 55 : 45 (52 %). 28 Since P,y-unsaturated cyclobutanones are known t o be susceptible t o attack by base with cy cleavage occurring at the position allyl t o the double b0nd,~2$2j~~7!28 structure 8 must be assigned t o the second isomer.

I

COOH 23

24

Final confirmation of structure 8 was given by the nmr spectrum (see Experimental Section), the most striking feature of which is the previously unreported long-range coupling (6.0 cps) between the H-1 and H-5 protons in the bicyclo[3.l.l]heptane systemO29 These protons occur as overlapping triplets centered at T 7.1 with J (H-7-exo, H-1) E J (H-7-ex0, H-5) = 6.0 cps, J (H-1, H-5) = 6.0 cps and J (H-I, H-7endo) = J (H-5, H-7-endo) E 0 cps (some broadening). The 7-exo proton appears as a multiplet centered at T 8.0 and superimposed with the C-2 methyl at 7 8.19 while the 7-endo proton appears as a doublet centered at T 8.30, J(7-endo,7-exo) = 9.0 cps. Irradiation of the bridgehead protons at T 7.1 converted the c-7-exo proton at T 8.1 into a doublet, J = 9.0 cps. Examination of Dreiding models of 8 reveals dihedral angles between the H-7-endo proton and the H-1 and H-5 protons of -85'; the result should be virtually no coupling between these protons as observed. The dihedral angle between the H-7-exo proton and the H-1 and

13

(27) J. de Pascual Teresa, H. Sanchez Bellido, and I . Sanchez Bellido,

The principal features of the mass, nmr, infrared, and ultraviolet spectra of compounds 2 and 8 were mark(23) The numbering system for bicyclic structures as accepted by Chemical Abstracts is used throughout this paper. (24) (a) H. Weitkamp and F. Korte, Tetrahedron, 20, 2125 (1964); (b) P. G . Gassman and F. V. Zalar, Tetrahedron Letfers, 3251 (1964). (25) J. J. Beereboom, J . Ani. Chern. Scc , 85, 3525 (1963). We have

assumed that the ketone isolated under the rather drastic conditions of heat and acid by Beereboom would equilibrate to the more stable cisfused structure 14. (26) The ketonc 13 is also produced from acid-catalyzed rearrangement of ketone 8. Chemical confirmation of structure will be reported latcr.11

Journal of'the American Chemical Society

I

COOH

Anales Real SOC.Espan. Fis. Quim. (Madrid), Ser. B, 58, 339 (1962). (28) The base cleavage of chrysanthenone and the ketone 8 and the

mechanistic implications of these cleavages will be discussed in another communication. (29) The spin coupling between the bridgehead protons of the bicyclo[2,2,l]heptanesystem has been recorded as 1.5 cps3°a while that of the bicyclo[l.l.l]pentane system is reported as 18 cps30b Spin coupling between the bridgehead protons of the bicyclo[2.l.l]hexane system is as yet u n r e c ~ r d e d . ~ ~ ~ ~ ~ (30) (a) R. W. King and P. E. Butler, Abstracts, 142nd National Meeting of the American Chemical Society, Atlantic City, N. J., Sept 1962, p 844; (b) K. B. Wiberg, G. M. Lampman, R. P. Ciula, D. s. Connor, P. Schertler, and J. Lavanish, Tefrahedron, 21, 2749 (1965); (c) J. Meinwald and A. Lewis, J . A m . Chern. Soc., 83, 2769 (1961); (d) K. B. Wiberg, B. R. Lowry, and B. J. Nist, ibid., 84, 1594 (1962).

1 89:15 1 July 19, 1967

3833

The observed coupling of 6.0 H-5 protons is -29". cps is that predicted from the Karplus equation. 3 1

8

The structure of piperitenone (15) was authenticated by comparison with a specimen of 15 prepared by the method of S c h i n ~and ~ ~the structure of isopiperitenone confirmed by comparison with an authentic sample prepared by the method of Schenck.33 Compound 11, described by the empirical formula CI0Hl40(molecular weight and elemental analyses) was a crystalline solid, mp 136-138", which possessed an unusually large optical rotation, [ a ]i-158'. ~ The infrared spectrum was that of a four-membered ring ketone or a strained five-membered ring, 5.69 p. The compound was not affected by hydrogen in the presence of 10 palladium-on-charcoal catalyst; vinyl protons were conspicuously absent from the nmr spectrum. The structure must therefore be tricyclic. The nmr spectrum disclosed only two methyl groups ( T 8.96 and 9.06) and these were noncoupled. That these methyl groups were not located geminate to each other was indicated by the infrared spectrum which showed only a single sharp peak at 7.25 p instead of the typical gem-dimethyl doublet. These data are consistent with only two structures, lla and b below. That isopiperitenone yields the same ketone on irradiation in cyclohexane (see Results) strongly supports this assignment. Careful examination of the infrared carbonyl absorption prompted us to assign structure lla to this ketone. Thus the 414 fused ketones 2534and 2Ci3j exhibit carbonyl absorptions of 5.61 and 5.63 p, respectively. In compari-

Hx HZ us.

HA 25, 5.61s

26, 5 . 6 3 ~

llb

son, the bicyclo[2.1. Ilhexanone (27)36absorbs at 5.67 p in the infrared and the closely related 3,8-cyclocamphor (28)37at 5.70 p comparable to the value for our tricyclic ketone (5.69 p). Absolute confirmation of structure was made by careful examination of the nmr spectrum and by double resonance experiments. The spectrum (in T values) of 11 revealed an eight-peak ABX multiplet consisting of a doublet at 7.50 (J = 3.0 cps), a quartet at 7.80 (J1 = 8.5 cps; Jz = 3.0 cps) and a doublet at 8.33 (J = 8.5 cps), a broad peak at 7.77, and a very complex AzBzC pattern in the region of 7.8-8.8. (The doublet at T 8.33 essentially bisects and does not appreciably overlap the A2B2Cpattern.) Irradiation of the doublet at T 7.50 with a frequency of 2701 cps converted the quartet at T 7.80 into a doublet, J = 8.5 cps, but had no effect upon the doublet at T 8.33. Irradiation of the quartet at T 7.80 with a frequency of 2668 or 2701 cps converted both peaks at T 8.33 and 7.50 into singlets. Examination of a molecular model of lla indicates that the four-membered ring is partially puckered and that the angle between the protons HA and Hx is approximately 80" ; thus the coupling constant JABshould be less than 1 cps. The angle HBHx is approximately 34" and should exhibit coupling in the order of 3-5 cps. The geminate coupling J A B in similar four-membered rings is reported to be in the order of 7-10 Since H A is located at the apex of the 7~ orbital of the ketone, this proton would be expected to occur at higher field than HB. Furthermore, the proton located exo to the two-carbon bridge is paramagnetically shielded relative to the protons endo to this bridge in similarly substituted bicyclo[2.1.13Thus the spectrum is completely consistent with structure lla with HA at T 8.33, HB at T 7.80, and Hx at T 7.50. In contrast the structure llb would not exhibit a simple ABX pattern as observed in the nmr spectrum of 11. Although a secondary coupling of 3 cps between the Hx,HB protons is possible, Hx would also show at least equal coupling with HZ. Thus the Hx signal would be expected to occur as a triplet (or quartet) which should become a doublet on irradiation of HB. The fact that the Hx signal degenerates into a singlet clearly precludes structure llb from consideration. That the ester product from irradiation of verbenone (1) or the ketone 14 in methanol was the nonconjugated isomer 3a or 4a was suggested from the infrared carbonyl absorption, 5.73 p, and the completely resolved nmr spectrum outlined below. 7.37 (t,J-7.Ocps)

28,5.70

lla, 5.69r (31) L. M. Jackman, "Applications of Nuclear Magnetic Resonance Spectroscopy in Organic Chemistry," Pergamon Press Inc., New York, N. Y . , 1959, p 87. (32) Ch. Balant, Ch. A. Vodoz, H. Kappeler, and H. Schinz, Helu. Chim. Acta, 34, 722 (1951). (33) G. 0. Schenck, 0 . - A . Neumuller, G. Ohloff, and S. Schroeter, Ann., 687, 26 (1965). (34) R. N. McDonald and C. E. Reineke, Tetrahedron Letters, 2739 (1966). ' (35) W. G. Dauben and R. M. Coates, J . A m . Chem. SOC.,86, 2490 (1964).

The cis structure 3a was confirmed by an unambiguous synthesis from cis-ocimene (29). Hydroboration (36) J. Meinwald and P. G. Gassman, ibid., 85, 57 (1963). (37) E. J. Corey, M. Ohno, S . W. Chon, and R. A. Schemer, ibid., 81, 6305 (1959).

Errnan 1 Ultraviolet Irradiation of Verbenone

3834

of 29 with disiamylborane, 3* subsequent Jones39 oxidation, and diazomethane methylation afforded the cis ester 3a. Repetition of the same sequence starting with trans-ocimene (30) produced the trans ester 4a which was well separated from 3a by gas chromatography.

29

3a

31

OH

b

dihedral angle (0') would require spin interaction of at least 8-10 cps. Absolute Configuration and Rotation of Products. The following arguments and experiments were employed to assess the absolute configuration and estimated optical rotations of compounds isolated or utilized in these reactions. The preparation of verbenone (1) from pinene inherently proceeds with optical retention. Since the pinene utilized in this work possessed a rotation of - 55' (chloroform) (-48 ", neat) approaching that of optically pure pinene,40 the rotation of -256" for verbenone approximates that of the optically pure compound. The absolute configuration of ( -)-verbenone is deduced from the known configuration of ( -)-pinene41 and is shown by structure 1 in Chart IV. Chart IV

32

30

4a

The structures of the three acetals 19,20, and 21 from irradiation of ketone 14 in methanol were described by the infrared spectra, which showed strong absorption in the 9-10-p region, characteristic of ether functions, and by their well-resolved nmr spectra outlined below.

14

1

L

,

7.3-8.1

2'0

19 8.31 6.92 4

f

33 .Lt

,

a

i

5.40 (J=1.0 CpS) -

M

6.80

CH3 8.91 7,1-7,&

H

CH3 8.71

34

7.35 ( J ,= J , =8.0; J , - 2 . 1 cps)

K

21

The C-7 methoxyl group is assigned the endo stereochemistry in 19 and the exo stereochemistry in 20 on the basis of the chemical shift of the methoxyl methyl protons and the C-7 protons of the two isomers. The C-7 proton in 20 lies over the x orbital of the C-2 carbon and should be diamagnetically shielded relative to the C-7 proton in 19. Indeed, the C-7 proton in 20 occurs at 0.06 ppm higher field than the same proton in 19. Similarly, the C-7 methoxyl protons in 19 occur at 0.13 ppni higher field than the C-7 methoxyl protons in 20, a consequence of the former being situated almost directly over the double bond in 19. The stereochemistry of the third acetal 21 is defined by the spin-coupling pattern of the C-8 proton a to the niethoxyl group. This proton, which occurs at 7 5.40, shows only small coupling ( < l . O cps) with the C-1 proton. Such small coupling would occur only if these protons were trans to each other. In this orientation the dihedral angle between the two protons , . , 107 '. If the protons were cis oriented the resulting is (38) H. C. Brown and G. Zweifel, J . Am. Chem. Soc., 82, 3223 (1960). (39) I